Peptidomic study of casein proteolysis in bovine mi lk by Lactobacillus casei PRA205 and Lactobacillus rhamnosus PRA331

Abstract Lactobacilli contain different cell envelope proteinases (CEPs) responsible for the hydrolysis of caseins and the release of various bioactive peptides. In this work, we explored the CEP activity of Lactobacillus casei PRA205 and Lactobacillus rhamnosus PRA331 whole cells towards β-, αS1-, κ- and αS2-caseins in bovine milk. Mass spectrometry analysis of fermented milk hydrolysates identified a total of 331 peptides, which were mainly derived from β-caseins (59.0 and 60.1% for PRA205 and PRA331, respectively). Analysis of αS1-casein (f1–23) cleavage site specificity congruently supports that Lb. casei PRA205 and Lb. rhamnosus PRA331 exhibited a mixed-type CEPI/III activity. PRA205 and PRA331 CEPs also showed cleavage site specificity toward β-casein. These CEPs cleaved the peptide bond preferentially when hydrophobic or negatively charged amino acids were present. 13.5% and 13.7% of peptides released by Lb. casei PRA205 and Lb. rhamnosus PRA331 CEPs were found to have 100% homology with previously identified bioactive peptides.


14
Food intake with the goal of improving human health is an ongoing focus for research. 15 Recommendations for the consumption of certain nutritious fermented foods date back to the 16 Hippocratic Corpus of Ancient Greece. The idea that lactic acid bacteria (LAB) fermenting milk are 17 responsible for enhancing health and delaying the human aging was first proposed by the Russian 18 scientist Elie Metchnikoff more than a century ago (Mackowiak, 2013). In the past years, research 19 has documented a wide range of health benefits exerted by dairy LAB, especially immune and 20 metabolic ones, and it is now focusing to decipher the microbial mechanisms underpinning these 21 health-promoting effects (Reid, 2015). 22 Some beneficial effects exerted by LAB are due to the generation of secondary metabolites 23 with health-promoting properties. The most important biogenic compounds in fermented milk are 24 the bioactive peptides released from caseins via the LAB proteolytic system. Biological activities 25 associated with such peptides include immunomodulatory, antibacterial, anti-hypertensive, 26 antioxidant, mineral binding, and opioid-like properties (Brown et al., 2017). In addition, dairy 27 LAB are auxotrophic for many amino acids and efficient casein breakdown is crucial to make LAB 28 competitive as dairy starters (S-LAB), as well as suitable to survive in ripened cheeses as non-29 starter LAB (NS-LAB) (Kunji, Mierau, Hagting, Poolman, & Konings, 1996). The amino acid 30 release also contributes to the aroma compound formation during cheese ripening and impacts 31 sensorial properties and consumer's acceptance of dairy foods (McSweeney & Sousa, 2000). 32 Cell envelope proteinases (CEPs) are large multi-domain proteins anchored to the cell wall 33 that catalyse the first step of hydrolysis of milk caseins into peptides. Different transport systems 34 then internalize these peptides into the cell, where they are further hydrolysed by numerous 35 intracellular peptidases (Savijoki, Ingmer, & Varmanen, 2006). Six different types of CEPs have 36 been described in several LAB species: PrtB from Lactobacillus delbrueckii subsp. bulgaricus 37 (Laloi, Atlan, Blanc, Gilbert, & Portalier, 1991); PrtH from Lactobacillus helveticus (Genay,Sadat,38 Gagnaire & Lortal, 2009); PrtL from Lactobacillus delbrueckii subsp. lactis (Villegas, Brown, ultra-filtrated fraction obtained from fermented milk as previously reported (Solieri et al., 2015). 114 The tripeptide 2-furanacryloyl-phenylalanylglycylglycine (FAPGG) was used as substrate assay 115 and the ACEi activity was calculated as percent of inhibition (ACEi%). 116 Three analytical replicates were run for each sample collected at the end of each 117 fermentation trial (carried out in triplicate) in all the assays. Chip enrichment column at a flow rate of 4 μL min -1 with a mobile phase consisting of 100% A 128 using a G1376A capillary pump. A flush volume of 2 μL and a flush-out factor of 5 were used. 129 After valve switching, a gradient elution was performed throughout the enrichment and analytical 130 columns at 500 nL min -1 using a G2226A nano pump. The gradient started at 0% B for 1 min, and 131 then linearly ramped up to 90% B in 70 min. The mobile phase composition was maintained at 90% 132 B for 15 min in order to wash both enrichment and analytical columns. The mass spectrometer was  For peptide identification, MS/MS spectra were converted to .mgf files and were then 136 searched against the Swiss-Prot database using Protein Prospector (http://prospector.ucsf.edu) and 137 parameters were considered: enzyme, none; peptide mass tolerance, ± 40 ppm; fragment mass 139 tolerance, ± 0.12 Da; variable modification, oxidation (M) and phosphorylation (ST); maximal 140 number of post-translational modifications permitted in a single peptide, 4. We considered only 141 peptides with a best expected value lower than 0.05 that corresponded to P<0.01. The assignment 142 process was complemented and validated by the manual inspection of MS/MS spectra. Three 143 replicates for each fermentation trial was injected in the mass spectrometer and only the peptides 144 present in at least two replicates were considered significant and included in the analysis. The identified peptides in milk samples were investigated for literature-identified bioactive peptides 147 using the BIOPEP database and the Milk Bioactive Peptide Database (MBPDB) (Minkiewicz,148 Dziuba, . Only peptides 149 with 100% homology to known functional peptides were considered as bioactive peptides.

150
The relative amount of the bioactive peptides was estimated by integrating the area under the peak 151 (AUP). AUP was measured from the extracted ion chromatograms (EIC) obtained for each peptide 152 and normalized to the peptide content of milk hydrolysates. The peptide content was determined at 153 the end of the fermentation trials by using the TNBS method as described in section 2.4 and 154 expressing the results as mg of leucine equivalent mL -1 155 . 156 2.7 Calculation of the cleavage specificity 157 The cleavage probability and positive or negative influence on the cleavage of an amino acid 158 in the P1 and P1' subsites were calculated according to Keyl (1992).

159
The subsite nomenclature was according to Schechter & Berger (1967) where the amino 160 acid residues are designated as P1 in the N-terminal direction (on left of the sequence) and Pl' in the 161 the subsite S1 in the enzyme active site, whereas the subsite P1' interact with the subsite S1' in the 163 enzyme active site. Therefore, the peptidic bond cleaved by the protease was defined as the P1-P1' 164 bond. We quantitatively analysed the influence of specific amino acid residues in position P1 or 165 P1'on the CEP cleavage probability.

166
If the amino acid residue A is in the position n (P1 or P1' subsite), the cleavage probability 167 of the P1-P1' bond will be: 11 (corresponding to the 6.2% of total identified peptides) and 8 peptides (corresponding to the 225 5.9% of total identified peptides), respectively. As expected, no significant proteolysis of whey 226 proteins was observed for both the strains. The Venn diagram (Figure 1) showed that 24.6 and 227 14.0% of peptides were specific for PRA205 and PRA331 milk hydrolysates, respectively. The 228 majority of the identified peptides were found in both the milk hydrolysates, suggesting that 229 PRA205 and PRA331 share a similar caseinolytic pattern.

230
CEPs are classified on the basis on their caseinolytic specificity (Kunji et al. 1996).

231
Typically, two CEPs have been identified: a PI-type, which preferentially hydrolyses β-casein, and a 232 PIII-type, which acts on αS1-, βand κ -caseins equally well (Pritchard & Coolbear, 1993 Amino acid sequence analysis of the identified peptides revealed that the cleavage sites were 278 not concentrated at the N-or C-terminus, but rather distributed throughout the entire β-casein 279 sequence for both the CEPs activities (Figure 3). Most of the proteinases previously described in  (Table 1). This CEP cleaved preferentially when the P1 position was 285 occupied by the hydrophobic amino acids M, L and F or the negatively charged amino acids Q and 286 N primarily, and by the polar un-charged amino acid E to a lesser extent. Table 1 (Monnet et al., 1992). Additionally, the residue P in both the P1 and P1' positons negatively 310 affected CEP cleavage in NCDO763 (Monnet et al., 1992). Similarly, the presence of a P residue 311 bound to one of the preferred cleaved amino acids prevented CEP from Lb. casei PRA205 to cut the 312 peptidic bond. For example, the preferentially cleaved amino acids Q and L formed seven and nine 313 peptidic bonds with the amino acid P, respectively, but no one of these bonds was cleaved by 314 PRA205 CEP (Figure 3). Finally, PRA205 CEP activity displayed the following two unique cleavage site-specificity pattern has never been described in lactobacilli.

318
As reported in Table 1 and  The peptides cleaved by PRA205 and PRA331 CEPs in milk hydrolysates were searched against 324 the general bioactive peptide database BIOPEP  and the milk bioactive 325 peptide database MBPDB , in order to find peptides which match sequences to 326 known bioactive peptides. Out of 331 identified peptides, 24 shared 100% homologies with 327 functional peptides previously reported to have various bioactivities ( showed in vivo antihypertensive activity in spontaneously hypertensive rat (Maeno, Yamamoto, & 341 Takano, 1996;. For all the 342 other identified bioactive peptides, the bioactivity was previously demonstrated with in vitro assays.

343
The physiological effects of bioactive peptides depend on their capability to arrive at the  (Cech, & Enke, 2000). Here, we compared the relative amount (expressed as AUP) of 381 the same peptide in two different hydrolysates coming from the same matrix (fermented milk), thus 382 we can exclude errors related to the extrinsic effect and assume that differences in peak intensity of 383 the same analyte accurately reflects relative differences in its abundance.

392
Nowadays, there is an increasing interest in developing novel dairy healthy products.     respectively. Please note that the amino acid C is not present in the sequence of β-casein, whereas W was omitted from the analysis since it occurs once in the β-casein sequence.
1  a One code letter was used for amino acid nomenclature. The amino acid C is not present in the sequence of β-casein, while the amino acid W was omitted from the analysis since it occurs once in the β-casein sequence. b The cleaved bonds are reported in Figure 3. c See materials and methods section for the calculation of the %P1 and %P1' cleavage probability.